DESCRIPTION: With increasing availability of more effective medicines and their use early in the course of disease, choice of therapy for JRA has become an important clinical issue. However, the proportion of patients who do not respond to a given therapy is sometimes substantial. For example, some 30-40% of patients do not benefit from methotrexate and a small proportion fail to respond to any therapy. It is likely that the capacity to respond or fail to respond to a drug(s) and the observed variation in toxicity are genetically determined and that methotrexate, one of the most widely used and effective treatments for JRA, may be taken as a paradigm in this regard. The Human Genome Project and its databases as well as a local JRA genomics database can be leverage to study methotrexate pharmacogenomics and used to test the hypothesis, that methotrexate resistance (or responsiveness) in JRA is a function of genetic diversity, as is overall susceptibility to the disease, will be tested as follows: Specific Aim 1: Determine if specific methotrexate responsiveness and/or toxicity genes exist in children with JRA using a candidate gene approach in a responder/non-responder study. Specific Aim 2: To replicate the association study findings by linkage analysis using transmission disequilibrium testing (TDT) in simplex and multiplex JRA families. Specific Aim 3: Modified genome-wide screen for methotrexate responsiveness and toxicity. Long-Term Goals: To match JRA patients to therapies most likely to give the best clinical effect with the least toxicity, i.e., have a good risk/benefit profile thus improving the quality of life for these clinically disabled patients. With respect to women and minorities, a substantial excess of this JRA population will be girls and all minorities will be included. A collaborative arrangement with the Pediatric Rheumatology service in Stanford will enrich the study population for Hispanics.